AI engineer and software developer in Sydney, building evaluated RAG systems, AI-code verification tools, local-first developer products, and production web systems.
Portfolio | Resume | LinkedIn | AI Posts | Email
- Automations and AI Specialist at Pretian Squared in North Sydney, building AI automation workflows, reusable delivery assets, and reviewable LLM-assisted systems.
- Master of Artificial Intelligence at the University of Technology Sydney, graduating December 2026.
- 3+ years across Pretian Squared, City Quokka, AI Camp, eden.pm, and IBM client work for Mercedes-Benz R&D and DBS Bank.
- Open to full-time AI engineering, software engineering, SDET, and ML engineering roles from December 2026.
| Project | What it demonstrates | Links |
|---|---|---|
| ISM CyberRAG | Evaluated cybersecurity RAG over 25 Australian ISM PDFs with control-aware chunking, hybrid pgvector/BM25 retrieval, cross-encoder reranking, refusal guardrails, RAGAS evaluation, and a live pipeline explorer. Faithfulness improved from 0.6834 to 0.8351 across the fixed benchmark. | Live Code Build note |
| Scout | Verification layer for AI-written code with specialist review agents, judge deduping, evidence graph, trace receipts, context budgeting, MCP server, and execution-gated patch scoring against a seven-bug seeded benchmark. | Live Code Build note |
| SR Terminal | Browser-based AI development environment with WebContainers for Node.js, Monaco, XTerm.js, and Phi-3 3.8B running locally through WebLLM/WebGPU. No backend is required for code execution or inference. | Live Code |
| SR Mesh | Local-first 3D knowledge graph using Transformers.js embeddings, a rule-based classifier, K-means++ clustering, IndexedDB persistence, and React Three Fiber visualization. Personal notes stay in the browser. | Live Code |
| MCP Prompt Library | AI development toolkit with 115+ curated prompts, an npm package, CLI, and MCP server exposing 15 tools for assistant-driven workflows. | Code |
| Role | Scope |
|---|---|
| Automations and AI Specialist, Pretian Squared | AI automation workflows, prompt design, evaluation patterns, guardrails, Python, TypeScript, and MCP-oriented tooling concepts. |
| Software Engineer, City Quokka | Full-stack CRM and CMS work with Next.js, TypeScript, Supabase, PostgreSQL, Docker, Playwright, and AWS. |
| SWE Intern, eden.pm | Serverless full-stack features with Next.js, Express.js, AWS Lambda, DynamoDB, Vercel, and CI/CD practices. |
| SDET, IBM for Mercedes-Benz R&D | Playwright and Cucumber automation, release-readiness testing, CI quality gates, and GenAI-assisted test-case workflows. |
| Application Developer, IBM for DBS Bank | Java 8, Spring Boot, REST APIs, JMeter performance work, production banking deployments, and distributed server environments. |
- Tech School: a public CS and AI learning hub with ELI5 explainers, coding quizzes, browser-based coding lessons, learning paths, comparison guides, and source-backed gotchas.
- AI Explorations: technical writing on LLMs, RAG systems, AI engineering, and behind-the-build notes for shipped projects.
- Portfolio archive: deployed AI systems, developer tools, web apps, and machine-readable profile surfaces for recruiters and AI agents.
- AI systems: RAG evaluation, retrieval, reranking, guardrails, MCP, agent workflows, local inference, and AI-code verification.
- Full-stack systems: Next.js, React, TypeScript, Python, FastAPI, Supabase/Postgres, pgvector, Docker, Vercel, Hugging Face Spaces.
- Quality: Playwright, Jest, Vitest, GitHub Actions, CI/CD, API validation, security-minded defaults, and evidence-backed technical writing.